[gmx-users] Problem with PME in LIE

[Justin Lemkul]

On 7/17/13 8:54 AM, Sainitin Donakonda wrote:
> Hi ,
>
> I want to use g_lie for my protein-drug complex to get binding energy ..i
> read some information that we need take care some issues if we used PME
> electrostatics..
>
> Indeed i have used PME in my simulation..
>
> Can any body explain which parameters to be taken care while running g_LIE
> and what is the isssue with PME ..
>

The long-range mesh terms are not decomposable in a pairwise manner.  LIE 
assumes that all of your interactions can be broken down from additive terms, 
and this is not the case with PME.

> and why we should use -rerun option beofre running LIE
>

If you evaluate the energies from a PME simulation, the result will be off.  If 
you use plain cutoffs for the dynamics, the simulation will probably be junk. 
So one solution is to run the simulation with PME to get reasonable dynamics, 
then post-process via mdrun -rerun while evaluating energies with plain cutoffs 
(which one would usually increase beyond normal values, IIRC, but I do not use 
LIE myself).

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441

==================================================